نموذج معدل للانحدار الخطي المتعدد باستخدام مرشحات تقليص الموجة الصغيرة "دراسة تطبيقية"

The researcher was interested as for his search in finding an efficient model with parameters of multiple-linear model and in diagnosis the outlier values through the use of analyzing the wavelet with the method of ordinary least squares (OLS) and some methods used for treating the outlier values and used in the treatment of outlier values and estimate the linear model such as least trimmed squares method (LTS) and M method .Therefore, this leads to revealing and treating the Outlier values and estimating the multi-linear model with wave shrinkage that involves the wave functions and the use of a method for estimating the level of cutting threshold including soft thresholding then comparing the efficiency of the estimates with ordinary and robust methods and fortified with some filters of wavelet with the application to patients having angina pectoris through proposing a method that works for inputting the outcomes of estimated values of robust methods as for the analysis of wavelet filter and estimating the best model of multiple linear regression model between the ratio of cholesterol in blood for the patients having angina pectoris who are enrolled in Specialized Center For Cardiology in Erbil in Kurdistan-Iraq and calculating some statistical standards such as (MSE,R2,F-Cal.) to compare types of wavelet in the multiple regression model and to reach the
3
best model . The research Concluded that the filters wavelet give the best results as for estimating the multiple linear regression model in comparison with the robust regression model in relation to statistical standards ( MSE, R2, F- Cal) and that the filter of wavelet Haar was the best in relation to the types and levels of wavelet Connected with estimating to the multiple linear regression model especially with LTS relying upon the standards (MSE,R2,MAPE,F-cal.,MSAE).